Dagstuhl Seminar 27061
Blending Knowledge Graphs and Process Mining for Procedural Knowledge Management
( Feb 07 – Feb 12, 2027 )
Permalink
Organizers
- Irene Celino (CEFRIEL - Milan, IT)
- Fajar Ekaputra (Wirtschaftsuniversität Wien, AT)
- Marco Montali (Free University of Bozen-Bolzano, IT)
- Han van der Aa (Universität Wien, AT)
Contact
- Andreas Dolzmann (for scientific matters)
- Jutka Gasiorowski (for administrative matters)
Processes lie at the heart of organizations. They structure operations across employees, systems, and technologies, generating outcomes of value to customers. At their core is procedural knowledge – the know-how required to perform tasks effectively. Unlike descriptive knowledge (knowing-what), procedural knowledge is often tacit, difficult to articulate, and scattered across unstructured sources such as manuals, regulations, slide decks, and meeting transcripts. This makes its systematic capture, integration, and exploitation an ongoing challenge.
Despite advances in areas such as Industry 4.0, process mining, and knowledge management, the full potential of procedural knowledge remains underused. Research communities have traditionally approached the problem from different angles: process mining and management on one side, semantic technologies and knowledge graphs on the other. Earlier attempts to integrate these perspectives were hampered by methodological burdens and limited adoption.
However, the emergence of large language models (LLMs) has transformed the landscape, making it feasible to unify procedural knowledge from heterogeneous and unstructured sources at scale, reducing barriers to semantic integration and enabling new opportunities. Therefore, this seminar comes at a key moment to bring together diverse expertise and jointly shape the next generation of procedural knowledge engineering.
This Dagstuhl Seminar will focus on three key goals:
- Establish a shared conceptual foundation: Clarify and reconcile how terms such as procedure, process, and workflow are understood across communities, to enable effective interdisciplinary collaboration.
- Identify new use cases enabled by LLMs: Identify and discuss new use cases for procedural knowledge engineering made possible by LLMs, especially those involving the integration of knowledge from diverse sources, such as process models and event data on the process side, as well as unstructured sources such as text, slide decks, and meeting transcripts.
- Investigate cross-community synergies: Investigate where and how methods and tools from the process mining and knowledge graph communities can be effectively combined to address these emerging use cases. How can we leverage the complementary strengths of both fields to co-develop the next generation of procedural knowledge solutions?
By bringing together researchers from diverse backgrounds, we aim to build a shared vision, foster cross-disciplinary collaboration, and pave the way for innovative applications of procedural knowledge engineering in research and practice.

Classification
- Artificial Intelligence
- Software Engineering
Keywords
- Procedural Knowledge
- Knowledge Graphs
- Process Mining
- Artificial Intelligence